"Multivariate time series datasets"

"The five datasets used in this study are GT-2011, PTB, MPI Roof, CHB-MIT, and ETT. The first four datasets are employed for the data imputation experiments, while the latter two are used for the data alignment experiments. A description of each dataset follows.GT-2011 is a dataset comprising measurements from 11 sensors, with a length of 36,733. It originates from a gas turbine and is intended for studying flue gas emissions, namely CO and NOx.PTB is a diagnostic electrocardiogram (ECG) dataset collected by the German Heart Center, containing 549 high-resolution 15-lead ECG recordings.MPI Roof contains meteorological observation data from the WS Beutenberg site in Germany, recorded from 2004 to 2023 at 10-minute intervals. It includes 21 variables such as temperature, air pressure, and humidity.CHB-MIT, collected by Boston Children\u2019s Hospital, contains electroencephalogram (EEG) recordings from pediatric subjects with refractory seizures. The recordings consist of 23 cases from 22 subjects; all signals are sampled at 256 samples per second.ETT (Electricity Transformer Temperature) is a dataset recording the load and oil temperature of two power transformers located at two power stations, collected at 15-minute intervals from July 2016 to July 2018.In the experiments, the adopted data segments contain 11 variables and 7400 samples for GT-2011, 15 variables and 10000 samples for PTB, 14 variables and 5000 samples for MPI Roof, 23 variables and 10000 samples for CHB-MIT, and 7 variables and 15000 samples for ETT."